From: "Alexei A. Frounze" Newsgroups: comp.os.msdos.djgpp Subject: Re: either I'm stupid or GCC once again has random behaviour Date: Fri, 5 Jan 2001 16:13:08 -0500 Lines: 26 Message-ID: <935dbh$90t5p$1@ID-57378.news.dfncis.de> References: <92o5uf$702n5$1 AT ID-57378 DOT news DOT dfncis DOT de> <978314359 DOT 13074 DOT 0 DOT nnrp-02 DOT 9e989cbe AT news DOT demon DOT co DOT uk> <92ophd$7kdsq$1 AT ID-57378 DOT news DOT dfncis DOT de> <8fd45t46ttftcmahol8erf6sipdjfs9v8v AT 4ax DOT com> <92tlcf$8accm$1 AT ID-57378 DOT news DOT dfncis DOT de> NNTP-Posting-Host: nas-204-59.rochester.navipath.net (64.20.204.59) X-Trace: fu-berlin.de 978729138 9467065 64.20.204.59 (16 [57378]) X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Reply-To: djgpp AT delorie DOT com does this really has to do anything with memset() ? -- Alexei A. Frounze alexfru [AT] chat [DOT] ru frounze [AT] ece [DOT] rochester [DOT] edu http://alexfru.chat.ru http://members.xoom.com/alexfru/ http://welcome.to/pmode/ "Damian Yerrick" wrote in message news:tcmb5t4e6r50eidnqsfp6kkkqrgsc5qta7 AT 4ax DOT com... > On Tue, 2 Jan 2001 17:41:07 -0500, "Alexei A. Frounze" > wrote: > > >Why unrolling? There is a perfect instruction REP STOSB/W/D. > >What do you mean? > > On modern x86 CPUs a loop of simple instructions (which can be split > among the U and V (and W in PII) pipelines) is actually _faster_ than > the string instructions (which are complex, run in microcode, and can > fit only in U). To fit code into the pipeline without time-consuming > branching, unrolling loops is a necessity.